*Balance Table for Behaviormean2 rep2 dem2 ppage educ ppgender ppincimp using balance.txt, ///over(condition) cat(Dissonant Balanced Consistent) excel replaceoneway rep conditiononeway dem2 conditiononeway ppage conditiononeway educ conditiononeway ppgender conditiononeway ppincimp conditionlabel define pid -1 "Republican"label define pid 0 "Independent", modifylabel define pid 1 "Democrat", addlabel values pid pid*Analyses **Independent Variablecondition-1 = dissonant0 = balanced1= consistent*this is an ordinal scale of dissonance-consonance*	3 = strong partisan & consistent*	2 = partisan & consisent*	1 = lean partisan & consistent*	-1 = lean partisan & dissonant*	-2 = partisan & dissonant*	-3 = strong partisan & dissonant*	-9 = all partisans in Balanced condition*			Notes: 	balanced condition =-9 because xi: automatically drops the lowest value, *					this ensures it is the reference value*					The "i.oconsist" code in xi: creates dummy variables for all values of oconsist*					and compares them against the reference value (in this case the balanced condition)*NOTE: the dependent variable for Figure 1 is perceived hostile/congenial bias*	where: *		-5=most biased against your position/party*		0=neutral*		5=most biased in favor of your position/party*	FIGURE 1 and TABLE A: MAIN HYPOTHESIS: PERCEIVED BIAS AND OBJECTIVITY	xi: reg bias consist dissonant balance if pid~=0, noconstant	estimates store mainbias	xi: reg bias consist dissonant balance if pid<0, noconstant	estimates store repsbias 		xi: reg bias consist dissonant balance if pid>0, noconstant	estimates store demsbias	coefplot mainbias , drop(rep ppage educ ppgender ppincimp) yline(0) xline(0) levels(90) 							///	ytitle(Perceived Bias) ylabel() xlabel(, angle(forty_five)) 			coefplot mainbias repsbias demsbias, drop() yline(0) xline(0) levels(90) 					///	ytitle(Experimental Condition) ylabel() xtitle(Perceived Hostile/Congenial Bias) 		///	xlabel(, angle(forty_five)) 		*note that the results are consistent when controlling for potential confounds in a single model (pid, ideo, educ)	xi: reg bias consist dissonant balance pid ideo educ, noconstant	coefplot, drop(_cons pid ideo educ) yline(0) xline(0) levels(90) 					///	ytitle(Perceived Bias) ylabel() xlabel(, angle(forty_five)) 	*TABLE A1: 	PERCEIVED BIAS AND OBJECTIVITY with LEANERS included	*creating condition variable to omit balanced from xi: reg because xi: automatically drops the lowest value, *		this ensures it is the reference value	gen condition2=condition	recode condition2 0=-10		*Basic, without leaners	xi: reg bias i.condition2 if pid~=0		outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace 		*Basic, with leaners		xi: reg bias i.condition2 if party~=0		outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace*TABLE A2: partisan interactions  	*Interaction Without leaners	xi: reg bias i.condition2*repdem if pid~=0		outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace		 	*Interaction, With leaners	xi: reg bias i.condition2*repdem if party~=0		outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace  		*FIGURE 2 (top panel) and TABLE B: SECOND MAIN HYPOTHESES part I: BELIEVABILITY of program as a whole	xi: reg Q11 consist dissonant balance if pid~=0, noconstant	estimates store mainbelieve	xi: reg Q11 consist dissonant balance if pid<0, noconstant	estimates store repsbelieve	xi: reg Q11 consist dissonant balance if pid>0, noconstant	estimates store demsbelieve	coefplot mainbelieve repsbelieve demsbelieve, drop(_cons pid ideo) yline(0) xline(5.5) levels(90) 					///	ytitle(Believability of Information and Analysis presented) ylabel() xlabel(1 2 3 4 5 6 7 8 9 10, angle(forty_five)) 	graph save believe "/Users/dikelly/Desktop/Research/Pol Comm Paper/Believe.gph", replace	(file /Users/dikelly/Desktop/Research/Pol Comm Paper/Believe.gph saved)*TABLE B1: Believability with party leaners	*Basic Believability, without leaners	xi: reg Q11 i.condition2 if pid~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace  		*Basic Believability, with leaners	xi: reg Q11 i.condition2 if party~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace		*TABLE B2: Believability with party interactions	*Believability Interaction Without leaners	xi: reg Q11 i.condition2*repdem if pid~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace	 	*Believability Interaction, with leaners	xi: reg Q11 i.condition2*repdem if party~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace	*	FIGURE 2 (bottom panel) and TABLE C SECOND MAIN HYPOTHESES part II: Informativeness 	xi: reg Q13 consist dissonant balance if pid~=0, noconstant	estimates store maininform	xi: reg Q13 consist dissonant balance if pid<0, noconstant	estimates store repsinform	xi: reg Q13 consist dissonant balance if pid>0, noconstant	estimates store demsinform	coefplot maininform repsinform demsinform, drop(_cons pid ideo) yline(0) xline(5.5) levels(90) 					///	ytitle(Informativeness) ylabel() xlabel(1 2 3 4 5 6 7 8 9 10, angle(forty_five)) 	graph combine believe inform, title("Figure 2: Perceived Credibility", size(medium)) col(1) iscale(*.8) imargin(medium)	///	ysize(10) xsize(5.5) name(Figure2, replace)		*TABLE C1: Informativeness with party leaners	*Basic Informativeness, without leaners	xi: reg Q13 i.condition2 if pid~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace  		*Basic Informativeness, with leaners	xi: reg Q13 i.condition2 if party~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace  *TABLE C2: Informativeness with party interactions	*Informativeness Interaction Without leaners	xi: reg Q13 i.condition2*repdem if pid~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace	 	*Informativeness Interaction, with leaners	xi: reg Q13 i.condition2*repdem if party~=0	outreg2 using basic, title(Percieved Bias) ctitle(fv2p,cens) se bdec(3) rdec(3)  replace																					